The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

CFV50c Images

darr

Well-known member
Hi again, Darr.
If there's a change that I'd be able to see the raw files and LCC shots for your above test, I'd be immensely grateful. It seems it's going to be very hard for me to demo a CMOS back for quite some time as my local distributor doesn't have a demo unit, just sale units for the next few months. I'd love to look at and process those shots myself to see what they look like.
TJV
Here is a link to a folder that has the RAW files for the +/-5, +/-10 and +/-15 files. These files will be deleted from the server in the near future so if you want them, please download them now.
 

Paul2660

Well-known member
Darr,

Many thanks for the raws. Does Phocus come in a trial version or does it happen to work like C1 which is free as long you are opening a Phase One raw file.

Paul
 

tjv

Active member
Many, many thanks, Darr! I'm downloading now and this will be extremely helpful!

Paul: In terms of Phocus, it's absolutely free to anyone who wants to use it. Just register and download from Hasselblad's website.
 

tjv

Active member
Thanks again, Darr.
I've had a little bit of time to look at the files and apply the scene calibration on each shot.
My initial impression is that the files look good, but I will probably need to choose a sensor that allows me to employ more movements. This is pretty gutting to me, because everything else about the package ticks all the boxes and them some... Hmmmmmm... Food for thought.
How are you finding working within the 10mm of shift limit you mentioned?
 

darr

Well-known member
Thanks again, Darr.
I've had a little bit of time to look at the files and apply the scene calibration on each shot.
My initial impression is that the files look good, but I will probably need to choose a sensor that allows me to employ more movements. This is pretty gutting to me, because everything else about the package ticks all the boxes and them some... Hmmmmmm... Food for thought.
How are you finding working within the 10mm of shift limit you mentioned?
So far it has not been a problem. I do like the color pallet coming out of the back better than my P45. But what works for me, may not work for you. I think to really know if this back would be right for you, you'll have to give it or the IQ250 a test drive.

It is a shame you cannot test or rent prior to purchase. As Paul stated earlier, it is frustrating we cannot rent digital backs without a ton of hassle or if it is even possible at all. The MFD business is a bit cut throat over sales and this is an example, I mean after all, I can rent a Mercedes or a BMW, but not a digital back. If the MFD business wanted to offer rentals, they would find profit in selling the insurance, but my guess is, it is not profitable enough for them, (or they may loose sales), so they push the sales and upgrade game.

Good luck with your decision.

Kind regards,
Darr
 
Last edited:

tjv

Active member
Hey all,

Darr, I hope you don't mind me posting these crops. If so, let me know and I'll take them down.

Just been looking at the edges of the raws after running the scene calibration. I turned off the sharpening tab in Phocus and exported the 10mm and 15mm shots for analysis. I used expired as full resolution, 16bit TIFF files. I've attached a crop of the edges here and am wondering what I'm seeing happening in the brick work on the 006 shot (the 15mm shot) and in the yellow paint on the 004 shot. To explain, it looks like there is some kind of aliasing / mazing happening. Am I right in thinking this? At first, I thought it might be a sharpening problem, which is why I turned it off. Any ideas on what it is and if it's related to any of the known problems when using movements on this CMOS sensor?

Thanks again Darr for sharing these and if you'd like me to remove the crops, let me know.

TJV

PS: Damn, for some reason the forum is downsampling the attachment. Any ideas how to fix this?
 

tjv

Active member
This time, now also with the DT library test of the IQ250 with 32 HR VS the IQ260 with 32 HR.
Both the CFV-50c and IQ250 can be seen to exhibit the same artefacts the more movements employed – in both cases starting from about 10mm / 12mm and getting worse the more you dial in. The IQ260 stays comparatively "clean", it seems. Perhaps I'm digging up old ground, here? If so, sorry. It's the first time I've been able to dig into these RAW files and am very grateful to be able to finally do so. If you click on the attachments, you can see them at 100%.
 
This time, now also with the DT library test of the IQ250 with 32 HR VS the IQ260 with 32 HR.
Both the CFV-50c and IQ250 can be seen to exhibit the same artefacts the more movements employed – in both cases starting from about 10mm / 12mm and getting worse the more you dial in. The IQ260 stays comparatively "clean", it seems. Perhaps I'm digging up old ground, here? If so, sorry. It's the first time I've been able to dig into these RAW files and am very grateful to be able to finally do so. If you click on the attachments, you can see them at 100%.
The mazing artifact may occur when you shoot with the CMOS back in landscape orientation and do rise/fall movements (or in portrait orientation if you do left/right movements). According to my experience up to date, it may only occur in the blue part in the LCC shot. It could ruin an image for some specific scenes if you are not aware of this issue and shift all the way up to hit the edge of the image circle. I do not observe any issue if the sensor is shifted along the longer edge (e.g. rise/fall in portrait orientation).

I can confirm that the safe movement for the HR23 is up to 8mm rise/fall on the IQ250 (landscape orientation), otherwise for interior shots some textures (high key or pure color) may get plagued by the mazing artifacts. Such amount of movement is right within the area of a fullframe 645 CMOS sensor, which suggest that the current lens may work if the IQ250/CFV-50C is enlarged to fullframe 645 CMOS sensor without an increase of pixel density. However I expect the current HR wide angles to fail if the fullframe 645 CMOS sensor increases pixel density.

I need to have my HR40 lens back to do more stress tests in this regard.
 

Paul2660

Well-known member
When the library test was done way back in early 2014, there was mention of some type of strange artifact with the IQ250 and the 60mm XL Schneider.

Looking at the crops of the Library, the IQ250 crop to me shows some strange artifacts but to me they are common when an image is zoomed into 200% or more. But I need to go back to those files if I still have them.

What I can't figure out is why the top image shows a definite cut out in the wood trim, follow the head of the white character, not sure if that's an angle or something else. Directly to the right, on the 250 shot, you see an indent in the light trim. Where as on the 260 shot, it's even. That has me stumped as if they are from different parts of the ceiling?

It's also odd that the 260 shot seems to much cleaner, much cleaner. When I looked at the files, I can remember just the opposite, i.e. the 250 images showed so much more detail in the shadows. I double checked the titles and they clearly point to the 250 on the top, and 260 on the bottom. Can't figure that one out either.

PS, I have really enjoyed this topic, lots of good info on the 50MP chip.

Paul
 

tjv

Active member
The think that disturbs me on the IQ250 shot above, which can also be seen on the 50c image in the bricks – although you have to look hard at this size – is the mazing artefacts in what should be flat or relatively flat textures. It's very prominent in the IQ250 shot and it also seems to introduce harsh and very ugly false detail around edges. It's likely that like Voidshatter has said that if you're aware of how and where it presents itself it can be worked around. Very interesting stuff.
 

torger

Active member
The mazing is a side effect of crosstalk. What happens is that you have leaks between pixels and since there are two greens and one red and blue in a 2x2 pixel group one green will get most leaks from red and the other from blue, and then the result is that the green pixel values separate even if it's a flat color.

This makes the demosaicer confused and it can start to produce detail that is not there, and you get mazing.

The degree of mazing depends on color, a color which is close to the camera's native neutral, ie red and blue levels are about the same, will cause little problems as the crosstalk will then not cause the greens to separate. But a color with a strong blue and weak red component or the other way around will cause mazing earlier on.

(Lumariver HDR's flatfield correction algorithm will compensate for this and level out the greens in areas of crosstalk, so you don't get any mazing, but you still get desaturation. You can also try RawTherapee with DCB demosaicer and enable green equiliberation and increase it until mazing disappears. If I get the raw test files I can post a demo.)

Note that the Sony sensor has offset microlenses, which is a trick to get better wide angle performance but it assumes the lens is centered. When shifting it will still help, but if you shift "too far" it will do the opposite and you get a sudden breakdown in performance and you can get mazing. But if you shift a bit less you get better performance in the same area of the image circle where you with the larger shift got worse. The first IQ250 library test posted here was made without that knowledge and therefore shows worse results than you get if you keep within limits.

Here's a picture showing microlens offset
http://www.sony.net/Products/di/en-us/translation_img/products/o4j5/feature/func_0_3.png
 
Last edited:

tjv

Active member
Thanks Anders.
It's really interesting to see these results for myself in the RAW files now, that's to some very generous forum members. The CMOS sensor seems to be brilliant at what it does best and with very moderate movements, but for my purpose –, regrettably as I really, really wanted to buy one – it would not give me the flexibility of moments that I regularly use. At this point in time I think the CCD 50mpx Kodak sensor is a winner for me and the H system is priced very well at the moment. It's interesting times in the MF market, for sure.
 

jlm

Workshop Member
posted elsewhere, but may be useful here as well:
just finished my L-plate for the V cameras; shown on the 205, but also fits the 500 and the SWC (adapter plate supplied which allows battery to be fitted without the hasselblad battery extender.)

details:

camera with L-plate mounts via Arca-style grooves in portrait or landscape mode. also fits RRS clamps (RRS pano clamp shown, but the quick clamps also work).

clears battery of CFV50-c digital back (also CFV39 and 16);
allows access to controls of 205 series cameras.
black anodized aluminum
index marks show lens centerlines
clamp is centered over the camera tripod hole for 205 and 500; moved forward to clear battery for the SWC

in limited production:

 
Last edited:

JCL212

New member
It looks great!!

Would you also show us pictures of the L-plate attaching an SWC?
I want to know how the battery adapter works.



posted elsewhere, but may be useful here as well:
just finished my L-plate for the V cameras; shown on the 205, but also fits the 500 and the SWC (adapter plate supplied which allows battery to be fitted without the hasselblad battery extender.)

details:

camera with L-plate mounts via Arca-style grooves in portrait or landscape mode. also fits RRS clamps (RRS pano clamp shown, but the quick clamps also work).

clears battery of CFV50-c digital back (also CFV39 and 16);
allows access to controls of 205 series cameras.
black anodized aluminum
index marks show lens centerlines
clamp is centered over the camera tripod hole for 205 and 500; moved forward to clear battery for the SWC

in limited production:
 
Last edited:

jlm

Workshop Member
on the SWC: the existing plate on the SWC gets removed and replaced (same screws) with the stainless steel plate shown. this move everything forward so the battery will clear.







and the L-plate itself, showing the cutaways (one for anti-rotation when screwed to the camera, one to clear the 205 controls and straps)
 
Top